Still under construction... you can grab a teddy bear image from the image library (which would by dynamically generated from a real collection in an actual implementation of this idea) and drag it over the textarea to see the textarea change to a div. Ideally, the words in the div would float around the absolutely positioned image, but I have no idea how (or if) you could even do that. As an alternative, I figured that maybe when you drop the image, it calculates which paragraph you dropped it into, as well as whether it should float to the left or right, and then would add an image tag just inside the textarea. Anyway, neither of those happen at the moment, but it does become a textarea, and calculating whether to float left or right would be easy. Parsing out the paragraph that the mouse is in would be only slightly harder, involving measuring the offset of the mouse with respect to the div's position, and comparing that to the line height of the text contained.